def test_gms_get_assertions_on_dataset_field():
"""lists all assertion urns including those which may not have executed"""
dataset_urn = make_dataset_urn("postgres", "foo")
field_urn = make_schema_field_urn(dataset_urn, "col1")
response = requests.get(
f"{GMS_ENDPOINT}/relationships?direction=INCOMING&urn={urllib.parse.quote(field_urn)}&types=Asserts"
)
response.raise_for_status()
data = response.json()
assert len(data["relationships"]) == 1
def fldUrn(tbl, fld):
return builder.make_schema_field_urn(datasetUrn(tbl), fld)
def get_assertions_with_results(
self,
validation_result_suite,
expectation_suite_name,
run_id,
payload,
datasets,
):
dataPlatformInstance = DataPlatformInstance(
platform=builder.make_data_platform_urn(GE_PLATFORM_NAME)
)
docs_link = None
if payload:
# process the payload
for action_names in payload.keys():
if payload[action_names]["class"] == "UpdateDataDocsAction":
data_docs_pages = payload[action_names]
for docs_link_key, docs_link_val in data_docs_pages.items():
if "file://" not in docs_link_val and docs_link_key != "class":
docs_link = docs_link_val
assertions_with_results = []
for result in validation_result_suite.results:
expectation_config = result["expectation_config"]
expectation_type = expectation_config["expectation_type"]
success = True if result["success"] else False
kwargs = {
k: v for k, v in expectation_config["kwargs"].items() if k != "batch_id"
}
result = result["result"]
assertion_datasets = [d["dataset_urn"] for d in datasets]
if len(datasets) == 1 and "column" in kwargs:
assertion_fields = [
builder.make_schema_field_urn(
datasets[0]["dataset_urn"], kwargs["column"]
)
]
else:
assertion_fields = None # type:ignore
# Be careful what fields to consider for creating assertion urn.
# Any change in fields below would lead to a new assertion
# FIXME - Currently, when using evaluation parameters, new assertion is
# created when runtime resolved kwargs are different,
# possibly for each validation run
assertionUrn = builder.make_assertion_urn(
builder.datahub_guid(
{
"platform": GE_PLATFORM_NAME,
"nativeType": expectation_type,
"nativeParameters": kwargs,
"dataset": assertion_datasets[0],
"fields": assertion_fields,
}
)
)
assertionInfo: AssertionInfo = self.get_assertion_info(
expectation_type,
kwargs,
assertion_datasets[0],
assertion_fields,
expectation_suite_name,
)
# TODO: Understand why their run time is incorrect.
run_time = run_id.run_time.astimezone(timezone.utc)
assertionResults = []
evaluation_parameters = (
{
k: convert_to_string(v)
for k, v in validation_result_suite.evaluation_parameters.items()
}
if validation_result_suite.evaluation_parameters
else None
)
nativeResults = {
k: convert_to_string(v)
for k, v in result.items()
if (
k
in [
"observed_value",
"partial_unexpected_list",
"partial_unexpected_counts",
"details",
]
and v
)
}
actualAggValue = (
result.get("observed_value")
if isinstance(result.get("observed_value"), (int, float))
else None
)
ds = datasets[0]
# https://docs.greatexpectations.io/docs/reference/expectations/result_format/
assertionResult = AssertionRunEvent(
timestampMillis=int(round(time.time() * 1000)),
assertionUrn=assertionUrn,
asserteeUrn=ds["dataset_urn"],
runId=run_time.strftime("%Y-%m-%dT%H:%M:%SZ"),
result=AssertionResult(
type=AssertionResultType.SUCCESS
if success
else AssertionResultType.FAILURE,
rowCount=result.get("element_count"),
missingCount=result.get("missing_count"),
unexpectedCount=result.get("unexpected_count"),
actualAggValue=actualAggValue,
externalUrl=docs_link,
nativeResults=nativeResults,
),
batchSpec=ds["batchSpec"],
status=AssertionRunStatus.COMPLETE,
runtimeContext=evaluation_parameters,
)
if ds.get("partitionSpec") is not None:
assertionResult.partitionSpec = ds.get("partitionSpec")
assertionResults.append(assertionResult)
assertions_with_results.append(
{
"assertionUrn": assertionUrn,
"assertionInfo": assertionInfo,
"assertionPlatform": dataPlatformInstance,
"assertionResults": assertionResults,
}
)
return assertions_with_results
def create_test_data(test_file):
assertion_urn = "urn:li:assertion:2d3b06a6e77e1f24adc9860a05ea089b"
dataset_urn = make_dataset_urn(platform="postgres", name="foo")
assertion_info = AssertionInfoClass(
type=AssertionTypeClass.DATASET,
customProperties={"suite_name": "demo_suite"},
datasetAssertion=DatasetAssertionInfoClass(
fields=[make_schema_field_urn(dataset_urn, "col1")],
dataset=dataset_urn,
scope=DatasetAssertionScopeClass.DATASET_COLUMN,
operator=AssertionStdOperatorClass.LESS_THAN,
nativeType="column_value_is_less_than",
aggregation=AssertionStdAggregation.IDENTITY,
nativeParameters={"max_value": "99"},
),
)
# The assertion definition
mcp1 = MetadataChangeProposalWrapper(
entityType="assertion",
changeType="UPSERT",
entityUrn=assertion_urn,
aspectName="assertionInfo",
aspect=assertion_info,
)
timestamps = [
1643794280350,
1643794280352,
1643794280354,
1643880726872,
1643880726874,
1643880726875,
]
msg_ids = []
# The assertion run event attached to the dataset
mcp2 = MetadataChangeProposalWrapper(
entityType="assertion",
entityUrn=assertion_urn,
changeType="UPSERT",
aspectName="assertionRunEvent",
aspect=AssertionRunEventClass(
timestampMillis=timestamps[0],
partitionSpec=PartitionSpecClass(
partition="[{'country': 'IN'}]",
type=PartitionTypeClass.PARTITION,
),
messageId=str(timestamps[0]),
assertionUrn=assertion_urn,
asserteeUrn=dataset_urn,
result=AssertionResultClass(
type=AssertionResultTypeClass.SUCCESS,
actualAggValue=90,
externalUrl="http://example.com/uuid1",
),
runId="uuid1",
status=AssertionRunStatusClass.COMPLETE,
),
)
mcp3 = MetadataChangeProposalWrapper(
entityType="assertion",
entityUrn=assertion_urn,
changeType="UPSERT",
aspectName="assertionRunEvent",
aspect=AssertionRunEventClass(
timestampMillis=timestamps[1],
partitionSpec=PartitionSpecClass(
partition="[{'country': 'US'}]",
type=PartitionTypeClass.PARTITION,
),
messageId=str(timestamps[1]),
assertionUrn=assertion_urn,
asserteeUrn=dataset_urn,
result=AssertionResultClass(
type=AssertionResultTypeClass.FAILURE,
actualAggValue=101,
externalUrl="http://example.com/uuid1",
),
runId="uuid1",
status=AssertionRunStatusClass.COMPLETE,
),
)
# Result of evaluating this assertion on the whole dataset
mcp4 = MetadataChangeProposalWrapper(
entityType="assertion",
entityUrn=assertion_urn,
changeType="UPSERT",
aspectName="assertionRunEvent",
aspect=AssertionRunEventClass(
timestampMillis=timestamps[2],
partitionSpec=PartitionSpecClass(
partition="FULL_TABLE_SNAPSHOT",
type=PartitionTypeClass.FULL_TABLE,
),
messageId=str(timestamps[2]),
assertionUrn=assertion_urn,
asserteeUrn=dataset_urn,
result=AssertionResultClass(
type=AssertionResultTypeClass.SUCCESS,
actualAggValue=93,
externalUrl="http://example.com/uuid1",
),
runId="uuid1",
status=AssertionRunStatusClass.COMPLETE,
),
)
mcp5 = MetadataChangeProposalWrapper(
entityType="assertion",
entityUrn=assertion_urn,
changeType="UPSERT",
aspectName="assertionRunEvent",
aspect=AssertionRunEventClass(
timestampMillis=timestamps[3],
partitionSpec=PartitionSpecClass(
partition="[{'country': 'IN'}]",
type=PartitionTypeClass.PARTITION,
),
messageId=str(timestamps[3]),
assertionUrn=assertion_urn,
asserteeUrn=dataset_urn,
result=AssertionResultClass(
type=AssertionResultTypeClass.SUCCESS,
actualAggValue=90,
externalUrl="http://example.com/uuid1",
),
runId="uuid1",
status=AssertionRunStatusClass.COMPLETE,
),
)
mcp6 = MetadataChangeProposalWrapper(
entityType="assertion",
entityUrn=assertion_urn,
changeType="UPSERT",
aspectName="assertionRunEvent",
aspect=AssertionRunEventClass(
timestampMillis=timestamps[4],
partitionSpec=PartitionSpecClass(
partition="[{'country': 'US'}]",
type=PartitionTypeClass.PARTITION,
),
messageId=str(timestamps[4]),
assertionUrn=assertion_urn,
asserteeUrn=dataset_urn,
result=AssertionResultClass(
type=AssertionResultTypeClass.FAILURE,
actualAggValue=101,
externalUrl="http://example.com/uuid1",
),
runId="uuid1",
status=AssertionRunStatusClass.COMPLETE,
),
)
# Result of evaluating this assertion on the whole dataset
mcp7 = MetadataChangeProposalWrapper(
entityType="assertion",
entityUrn=assertion_urn,
changeType="UPSERT",
aspectName="assertionRunEvent",
aspect=AssertionRunEventClass(
timestampMillis=timestamps[5],
partitionSpec=PartitionSpecClass(
partition="FULL_TABLE_SNAPSHOT",
type=PartitionTypeClass.FULL_TABLE,
),
messageId=str(timestamps[5]),
assertionUrn=assertion_urn,
asserteeUrn=dataset_urn,
result=AssertionResultClass(
type=AssertionResultTypeClass.SUCCESS,
actualAggValue=93,
externalUrl="http://example.com/uuid1",
),
runId="uuid1",
status=AssertionRunStatusClass.COMPLETE,
),
)
fileSink: FileSink = FileSink.create(
FileSinkConfig(filename=test_file),
ctx=PipelineContext(run_id="test-file"))
for mcp in [mcp1, mcp2, mcp3, mcp4, mcp5, mcp6, mcp7]:
fileSink.write_record_async(RecordEnvelope(record=mcp, metadata={}),
write_callback=NoopWriteCallback())
fileSink.close()